PENSA+: Extended Follow-up of a Multimodal Lifestyle Intervention to Prevent Cognitive Decline in APOE-ε4 Carriers

Global cognitive performance assessed using the Preclinical Alzheimer Cognitive Composite for executive function, a composite score derived from six neuropsychological tests: Montreal Cognitive Assessment (score 0-30, higher scores indicate better cognition), Free and Cued Selective Reminding Test (score 0-48, higher scores indicate better episodic memory), Logical Memory Delayed Recall from the Wechsler Memory Scale (score 0-48, higher scores indicate better delayed verbal memory), WAIS-IV Coding (higher scores indicate better processing speed), Stroop Color-Word Test Interference score (higher scores indicate better inhibitory control and executive function), and Five Digit Test (higher scores indicate better executive functioning and cognitive flexibility). Individual test scores are converted to standardized z-scores based on the baseline mean and standard deviation of the study cohort, and the PACC-exe score is calculated as the arithmetic mean of these z-scores.

Incidence of Mild Cognitive Impairment determined according to clinical diagnostic criteria based on neuropsychological evaluation and clinical assessment performed at the 5-year follow-up visit.

Diagnosis will require: (1) evidence of objective cognitive impairment on standardized neuropsychological assessment; (2) reported cognitive decline from previous functioning by the participant, an informant, or documented longitudinal assessment; (3) preserved independence in activities of daily living (impaired performance will be defined as below the 10th percentile, scaled score <7, or ≥1.3 SD below age- and education-adjusted normative means); and (4) absence of dementia. Domain-specific impairment will be defined as low performance in ≥3 of 5 memory measures, ≥3 of 5 executive function measures, or ≥2 of 2 language measures. Final diagnosis will be established by a neurologist based on the integrated evaluation of neuropsychological, clinical, functional, and behavioral information.

Dementia risk assessed using the Lifestyle for Brain Health (LIBRA) Index, a weighted composite score of modifiable risk and protective factors for dementia. Lower scores indicate lower estimated dementia risk.

The dementia risk assessed with the LIBRA Index is a weighted composite score of 12 modifiable risk and protective factors for dementia. These modifiable risks are the following:

• Coronary heart disease
• Chronic kidney disease
• Hypertension
• Hypercholesterolemia
• Diabetes
• Depression
• High cognitive activity
• Obesity
• Low/moderate alcohol use
• Physical inactivity
• Smoking
• Healthy diet

Longitudinal changes in cognitive domains (memory and executive function scores, PMCID: PMC12413734).

Diet will be assessed through Mediterranean Diet Adherence Screener (MEDAS-14). The MEDAS-14 is a 14-item questionnaire for the assessment of Mediterranean diet compliance. It has been developed and validated by researchers of the PREDIMED study (PMCID: PMC8710807). The maximum possible score is 14 points.

This questionnaire allows the participants to be classified according to their adherence to the Mediterranean diet, being appropriate if the score is ≥10 or not appropriate if the score is <10.

Physical activity is measure through the Short Minnesota Leisure Time Activities Questionnaire (VREM). The VREM questionnaire asks for information on six activities carried out in the last month (or usual month) in order to find out the level of PA performed: walking, gardening in the garden, doing sport or dancing, climbing stairs, shopping on foot, and cleaning the house. cleaning the house (PMID: 23223762). According to the number of minutes spent on each activity and the frequency they are carried out over the course of a year, the energy expenditure is estimated for 14 days, expressed in Metabolic Equivalents (METs) per minute per 14 days (MET-min-14 days). The VREM makes it possible to distinguish between very active (more than 5,000 MET-min-14 days), active (between 3,000 and 4,999), moderately active (between 1,250 and 2,999), and sedentary (less than 1,250).

Quality of life is measured through the World Health Organization Quality of Life (WHOQOL)-BREF. The WHOQOL-BREF comprises 26 items, categorized into four domains: physical health (7 items), psychological well-being (6 items), social connections (3 items), and environmental health (8 items) (PMID: 15085902). Additionally, it includes items addressing the overall quality of life and general health. Each item in the WHOQOL-BREF is assessed on a five-point ordinal scale. These scores are then linearly transformed to a 0-100 scale. The physical health domain encompasses aspects such as mobility, daily activities, functional capacity, energy, pain, and sleep. Within the psychological domain, assessments cover self-image, negative thoughts, positive attitudes, self-esteem, mentality, learning ability, memory concentration, religion, and mental status. The social relationships domain explores personal relationships, social support, and aspects of one's sex life.

Physical fitness will be assessed using the Senior Fitness Test (SFT), a standardized battery that evaluates overall physical fitness in older adults, including muscle strength, aerobic endurance, flexibility, agility, and dynamic balance. Higher scores in the SFT indicate better physical fitness.

Physical Fitness will be assessed with the Grip Strength Test. For this test, the best result from 3 trials for each hand will be recorded and summed.

Change and trajectory in plasma AD biomarkers and neuroimaging measures of brain structure and function. Specifically, plasma AD biomarkers include the concentration levels of phosphorylated tau (p-tau217 and p-tau181), neurofilament light chain (NfL), and glial fibrillary acidic protein (GFAP), all expressed in pg/mL, as well as the amyloid-β 42/40 ratio. Simultaneously, neuroimaging markers derived from structural MRI include hippocampal, lateral ventricular volumes, and White Matter Hyperintensities (WMH) - all normalized to total intracranial volume (TIV)-, and a composite measure of cortical thickness in "AD-signature" regions (PMID: 21490323).

Incremental cost-effectiveness ratios (ICERs) expressed as cost per unit improvement in global cognitive performance (PACC-exe), and the Incremental Cost-Utility Ratio (ICUR), expressed as the cost per Quality-Adjusted Life Year (QALY) gained based on EQ-5D-5L data.

Incidence of all-cause dementia and AD identified via validated Electronic Health Record (EHR) algorithms.

Time-to-event methods will be used where feasible to analyze longer-term incidence outcomes. Kaplan-Meier curves and Cox proportional hazards models, or alternative parametric/survival models if proportional hazards assumptions are not met, will be used to compare intervention groups with adjustment for baseline covariates. If event dates are not sufficiently precise or follow-up intervals are irregular, discrete-time survival models or regression models for cumulative incidence may be used instead.

Longer-term healthcare utilization will be summarized descriptively and analyzed using count models (e.g., Poisson or negative binomial regression) or generalized estimating equations, as appropriate to the distribution of the data.

Concordance between participants and their study partners regarding Mediterranean diet adherence, measured through the Mediterranean Diet Adherence Screener (MEDAS-14).

Concordance between participants and their study partners regarding physical activity adherence, measured through the Short Minnesota Leisure Time Activities Questionnaire (VREM).

Concordance between participants and their study partners regarding cognitive engagement, which is measured through the Lifestyle Activities Questionnaire (LAQ). Cognitive and social activity will be measured using a questionnaire that results from the translation and adaptation of the LAQ. The LAQ is used to assess a participant's engagement in various daily activities with different cognitive demands, such as reading, playing musical instruments, or attending social gatherings. This questionnaire allows the quantification of both the frequency and the variety of these activities. A greater variety of participation in activities, as assessed by LAQ, was associated with reduced cognitive impairment (PMCID: PMC3508669).